[rfc]: Fix gdb.base/gcore.exp

[rfc]: Fix gdb.base/gcore.exp

[Markus Deuling]

[-- Attachment #1: Type: text/plain, Size: 289 bytes --]

Hi,

gdb.base/gcore.exp quits with a tcl error because of an uninitialized variable.
This patch initializes it in every case.

Is this ok ?

ChangeLog:

	* gdb.base/gcore.exp: Initialize variable
	core_supported.


-- 
Markus Deuling
GNU Toolchain for Linux on Cell BE
deuling@de.ibm.com


[-- Attachment #2: diff-fix-gcore-exp --]
[-- Type: text/plain, Size: 534 bytes --]

diff -urN src/gdb/testsuite/gdb.base/gcore.exp dev/gdb/testsuite/gdb.base/gcore.exp
--- src/gdb/testsuite/gdb.base/gcore.exp	2007-01-09 18:59:11.000000000 +0100
+++ dev/gdb/testsuite/gdb.base/gcore.exp	2007-05-03 19:04:07.000000000 +0200
@@ -108,6 +108,7 @@
 
 set escapedfilename [string_to_regexp ${objdir}/${subdir}/gcore.test]
 
+set core_supported 0
 gdb_test_multiple "gcore ${objdir}/${subdir}/gcore.test" \
 	"save a corefile" \
 {
@@ -123,7 +124,6 @@
   }
 }
 
-global core_supported
 if {!$core_supported} {
   return -1
 }

Re: [rfc]: Fix gdb.base/gcore.exp

[Ulrich Weigand]

Markus Deuling wrote:

> gdb.base/gcore.exp quits with a tcl error because of an uninitialized variable.
> This patch initializes it in every case.

The test case currently has:

gdb_test_multiple "gcore ${objdir}/${subdir}/gcore.test" \
        "save a corefile" \
{
  -re "Saved corefile ${escapedfilename}\[\r\n\]+$gdb_prompt $" {
    pass "save a corefile"
    global core_supported
    set core_supported 1
  }
  -re "Can't create a corefile\[\r\n\]+$gdb_prompt $" {
    unsupported "save a corefile"
    global core_supported
    set core_supported 0
  }
}


What output do you get that runs into neither of the two
tested patterns?

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com

Re: [rfc]: Fix gdb.base/gcore.exp

[Daniel Jacobowitz]

On Tue, May 08, 2007 at 08:26:54PM +0200, Ulrich Weigand wrote:
> What output do you get that runs into neither of the two
> tested patterns?

FWIW, while Alan had core files broken because of the
segment-in-section change, I encountered the same TCL error.

-- 
Daniel Jacobowitz
CodeSourcery

Re: [rfc]: Fix gdb.base/gcore.exp

[Markus Deuling]

Daniel Jacobowitz wrote:
> On Tue, May 08, 2007 at 08:26:54PM +0200, Ulrich Weigand wrote:
>> What output do you get that runs into neither of the two
>> tested patterns?
> 
> FWIW, while Alan had core files broken because of the
> segment-in-section change, I encountered the same TCL error.
> 

I cannot reproduce it with current head :-( There was an error in
creating that core file. 

For my opinion that variable should be initialized in every case. The testcase
should return a FAIL if GDB's output doesn't match the reg exp but not exit with an error.

<snip>
global core_supported
if {!$core_supported} {
  return -1
}
</snip>

However, this testcase runs without error now. Thanks for your effort. 

-- 
Markus Deuling
GNU Toolchain for Linux on Cell BE
deuling@de.ibm.com

Re: [rfc]: Fix gdb.base/gcore.exp

[Daniel Jacobowitz]

On Thu, May 03, 2007 at 07:08:23PM +0200, Markus Deuling wrote:
> Hi,
> 
> gdb.base/gcore.exp quits with a tcl error because of an uninitialized variable.
> This patch initializes it in every case.
> 
> Is this ok ?
> 
> ChangeLog:
> 
> 	* gdb.base/gcore.exp: Initialize variable 
> 	core_supported.

Yes, this is OK.  (No need to wrap in the middle of the log entry -
you've got plenty of columns :-)

-- 
Daniel Jacobowitz
CodeSourcery

Re: [rfc]: Fix gdb.base/gcore.exp

[Ulrich Weigand]

Daniel Jacobowitz wrote:
> On Thu, May 03, 2007 at 07:08:23PM +0200, Markus Deuling wrote:
> > 
> > 	* gdb.base/gcore.exp: Initialize variable 
> > 	core_supported.
> 
> Yes, this is OK.  (No need to wrap in the middle of the log entry -
> you've got plenty of columns :-)

I've committed this now, thanks.

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com